An adaptive complexion extraction method based on fuzzy entropy and fuzzy inference

This paper presents a novel adaptive method for the complexion extraction. Under the framework of the fuzzy entropy theory, our proposed method can automatically select the best segmentation points (the optimal thresholds) between the skin regions and background regions based on the skin color information. Furthermore, utilizing the fuzzy inference, the extracted optimal thresholds can be used to separate the skin color regions from the picture. The proposed extraction scheme is compared to the method of extracting the skin area with a fixed threshold, and the experimental results show that our method has fewer errors and more accurate.